A hybrid surrogate and simulation-based framework for efficient CapEx/OpEx optimization in complex chemical plants

Process Intensification aims at the economic and operational efficiency of chemical processes by emphasizing energy integration, unit size reduction, and cost minimization. The optimality of intensified solutions is typically assessed using Process Simulators, especially for complex chemical processes. These tools offer limited reliability and flexibility for the optimization of Capital Expenditures (CapEx), thus restricting their scope to Operational Expenditures (OpEx). As an alternative, external software is commonly required for simulation-based (SIM-OPT) and surrogate-based (SUR-OPT) CapEx/OpEx optimization. This work introduces a framework to select the most efficient optimization methodology based on simulation computational complexity. In addition, it presents a novel methodology (MIX-OPT) providing an efficient trade-off between optimization speed and accuracy. These three approaches were employed to optimize a complex biogas-to-methanol plant. Results showed that SIM-OPT achieved the greatest reduction in the Payback Period (PBP) of the plant (9.28%) with highest computational demand (984 min), SUR-OPT had the shortest computational time (717 min) with moderate PBP reduction (7.89%), and MIX-OPT reached a compromise with a PBP reduction of 8.24% in 884 min. The proposed framework demonstrated that simple simulations benefit from SIM-OPT, complex ones from SUR-OPT, and a wide range of simulations in the middle from the novel MIX-OPT approach. • Presented a comprehensive framework for selecting the most efficient simultaneous CapEx/OpEx optimization approach based on simulation complexity. • The framework + compares and integrates three different simulation and surrogate-based optimization methodologies to optimize complex production processes. • A novel optimization methodology combining surrogate modeling and simulation-based optimization is proposed. This approach demonstrated a 10% reduction in computational effort compared to traditional simulation-based optimization, while maintaining a high level of precision, with a 1% deviation from the optimum. • The framework was validated on a complex methanol synthesis simulation by analyzing the strengths and weaknesses of the three optimization methodologies. • Both Operational and Dimensioning parameters can be defined as degrees of freedom for the optimization problem.